Study of Selfish Gene Algorithm based automatic stimuli generation

Based on analyzing the evolution mechanism of SGA (Selfish Gene Algorithm), an implementation methodology on automatic stimuli generation and stimuli optimization in digital IC (Integrate Circuit) verification was studied in this paper. Basic items of SGA in the area of automatic stimuli generation, including individual definitions, the fitness function definition and the method of control structure extraction from RTL (Register Transfer Level) model were discussed. A two-stage-extract method was put forward, by which the traditional extraction process was divided into two orthogonal stages to reduce the complexity of the algorithm and improve the simulation efficiency of the generated stimuli. Experimental results show that the method is more efficient than traditional approaches in automatic simulation stimuli generation and stimuli optimization.